What if a relatively small change in rainfall could trigger an order-of-magnitude change in erosion? In the southern Central Andes, erosion appears to pulse in step with a four-hundred-thousand-year orbital rhythm. Climate has long been hypothesized to modify both the pattern and magnitude of erosion in mountainous landscapes, affecting morphology, deformation, and potentially global carbon and nutrient cycles.

But geologic evidence for a direct climatic control on erosion has remained ambiguous because high-resolution, long-term terrestrial records and suitable field sites are scarce. This study provides direct terrestrial field evidence for long-term synchrony between erosion rates and Milankovitch-driven, four-hundred-thousand-year eccentricity cycles, using a Plio-Pleistocene cosmogenic radionuclide paleo-erosion record from the southern Central Andes.

Combined with the CLIMBER-2 climate model, the observed coupling is consistent with relatively modest precipitation fluctuations producing synchronous and nonlinear erosion responses. The hypothesis that climate exerts a first-order control on landscape form and function through physical and chemical erosion dates back more than a century, but quantifying that influence in a particular landscape has proven difficult.

Numerical models provide rationales for precipitation affecting deformation and sediment flux, while empirical evidence remains equivocal, with evidence both for and against strong coupling between climate, erosion, and tectonics. Part of the ambiguity comes from temporal and spatial mismatches in the datasets and proxies used to infer these links.

Many interpretations compare present-day precipitation patterns with erosion or tectonic rates measured over timescales from ten squared to ten to the sixth years. High-quality erosion-rate data from terrestrial sediment records can reveal how landscapes respond to climatic and tectonic perturbations across different forcing frequencies.

Milankovitch cycles include precession at about twenty-three thousand years, obliquity at about forty-one thousand years, and eccentricity at about one hundred thousand and four hundred thousand years. These orbital periodicities change the seasonal, latitudinal distribution of solar insolation and are recognized as global climate drivers in marine and atmospheric records.

Yet linking sediment flux or erosion rates to these climate drivers over multiple cycles has remained largely theoretical, especially in active tectonic settings. The study presents forty-nine beryllium ten and twenty-three aluminum twenty-six cosmogenic radionuclide-derived paleo-erosion rates from the Río Iruya canyon in northwest Argentina, spanning six to one point eight million years.

This approach is similar to calculating present-day erosion rates from cosmogenic radionuclide concentrations in river sediments, which quantifies the combined signal from physical erosion and chemical weathering in the contributing catchment. Cosmogenically produced nucleons attenuate exponentially with depth through the upper few meters of rock and regolith, where they produce beryllium ten and aluminum twenty-six in quartz and accumulate over time.

Because radionuclide concentration is inversely proportional to erosion rate, concentrations in surface samples can be converted into erosion rates. The Río Iruya canyon is an ideal location because its depositional foreland basin lies close to a steep source area, less than one hundred kilometers away, minimizing prolonged sediment storage during transport.

Rapid foreland-basin deposition, about one meter per thousand years, suggests that concentrations at deposition were minimally affected by post-depositional nuclide accumulation during burial. A flood-diversion project completed in eighteen sixty-five rerouted and shortened the mainstem Río Iruya, and subsequent incision of about one hundred meters exposed more than eight kilometers of continuous Miocene-to-Pleistocene foreland sediments.

Rapid canyon exhumation, continuous high-resolution magnetochronology, and a simple pre-burial transport history permit a robust record of erosion-rate variability while minimizing post-depositional sources of beryllium ten and aluminum twenty-six. Figure one situates the Río Iruya study area on the eastern flank of the northern Andes, where the South American low-level jet brings rainfall toward the steep Eastern Cordillera and the Puna Plateau.

Panel b compares modern peak mean annual rainfall with CLIMBER-two model values, highlighting a transitional zone between about eighteen and twenty-five degrees south. Panels c and d then locate the watershed, channel diversion, and cosmogenic-radionuclide sampling sites spanning roughly six to one-point-eight million years, establishing the geographic and climatic context for reconstructing erosion.

The modern Río Iruya region has a semi-arid subtropical climate, with mean annual precipitation varying from less than zero point five meters per year on the arid Puna Plateau to about two point five meters per year on high-relief, east-facing mountain slopes. Nearly all precipitation comes from easterly or northeasterly sources because the Puna Plateau and hyper-arid Atacama Desert form a topographic barrier immediately to the west.

Roughly eighty percent of precipitation falls during the austral summer, between November and February, when easterly winds bring Atlantic moisture and fuel the South American Monsoon System over Brazil. Some of that moisture reaches the Río Iruya watershed through the South American Low Level jet, and precipitation is sensitive to the strength and position of the monsoon and that jet.

Very little is known about terrestrial Andean climate during the Pliocene-Pleistocene because high-resolution lacustrine and glacial records are lacking. The study therefore uses the low-resolution, intermediate-complexity CLIMBER-2 model to assess climatic forcing.

CLIMBER-2 has a coarse spatial resolution, with ten-degree latitudinal resolution and about fifty-one-degree longitudinal resolution, and it is insensitive to regional Andean orographic forcings such as the South American Low Level jet. Even with those limitations, the model provides a broad understanding of the magnitude and frequency of hemispheric moisture transport to the study region during the Plio-Pleistocene.

Paleo-erosion rates from approximately six to one point eight million years can be grouped into three periods: six to four million years, four to two point four million years, and two point four to one point eight million years. The transitions are roughly contemporaneous with previously documented tectonic and sediment-provenance shifts at about four point zero and two point four million years.

All reported paleo-erosion concentrations and rates were corrected for exhumation, decay, and burial components. The burial correction was the largest modification outside decay, ranging from one to thirty-six percent, with a mean of ten percent, while exhumation modification was negligible at less than zero point zero five percent.

At about four million years, beryllium ten concentrations record the onset of pronounced periodicity in erosion rates from about four to two point four million years. This interval contains up to three complete paleo-erosion cycles, with greater-than-tenfold rate changes from about zero point zero two to zero point two three millimeters per year.

Those changes are largely in phase with predicted Milankovitch long-eccentricity cycles at the four-hundred-thousand-year frequency. Figure four aligns global climate indicators with Río Iruya erosion evidence from roughly six to one point eight million years ago. It shows benthic oxygen isotopes, austral-summer insolation, corrected ten-Be and twenty-six-Al concentrations, and CLIMBER-two modeled precipitation, including filtered twenty-three-, forty-one-, and four-hundred-thousand-year components.

The key point is that from about four to two point four million years ago, the erosion record develops pronounced cycles that are broadly in phase with the four-hundred-thousand-year eccentricity signal, linking landscape change to climate variability. Figure seven compares CLIMBER-2 wavelet and power spectra for precipitation and temperature in the study region.

Precipitation shows strong twenty-three-kyr periodicities from four point five to three Ma, followed by strong forty-one-kyr periodicities from three to two Ma. Temperature is dominated more consistently by four-hundred-kyr periodicities, while one-hundred-kyr and forty-one-kyr components increase in power.

This matters because it distinguishes precipitation’s shorter orbital pacing from temperature’s longer-cycle signal, informing possible explanations for erosion-rate cyclicity. The CLIMBER-2 results suggest that precipitation oscillations are dominated by twenty-three-thousand-year precession and forty-one-thousand-year obliquity bands during this period.

At the same time, four-hundred-thousand-year eccentricity cycles strongly modulate the amplitude of the twenty-three-thousand-year precipitation cycles. When eccentricity is high, precession produces about twenty-percent swings in summer insolation at twenty-three degrees south; during low eccentricity, those cycles are damped to about two percent.

The final erosion-rate period, from about two point four to one point eight million years, departs from the four-hundred-thousand-year cyclicity. Instead, erosion rates show a large increase and fluctuate around a mean consistent with the modern rate of about zero point thirty-four millimeters per year.

Figure six tests Pearson correlations between erosion rates and three climate measures across three time periods. The strongest reported relationships occur from two point four nine to three point eight million years ago: erosion correlates with four-hundred-kiloyear eccentricity at zero point eight zero, and with CLIMBER-two precipitation at zero point four nine.

These results matter because they support the authors’ interpretation that orbitally paced precipitation changes, likely linked to insolation, helped drive erosion-rate cycles between roughly four and two point four million years ago. The most striking result is the correlation between erosion rate and the four-hundred-thousand-year amplitude modulation of twenty-three-thousand-year summer-insolation oscillations from about four point zero to two point four million years.

Because observed erosion cycles are synchronous with the amplitude of precipitation changes predicted by CLIMBER-2, the study hypothesizes that erosion-rate changes during this period were driven by summer monsoon precipitation. The preservation of orbitally paced erosion cycles during this limited interval, about four to two point four million years, also depended on favorable tectonic and geomorphic conditions in the watershed.

The clear cyclicity began at about four million years, alongside a major sediment-provenance shift interpreted to record a more restricted sediment source and possible sediment addition from the interior margin of the Puna Plateau. Figure eight links reconstructed erosion to orbital climate forcing.

Panel A shows ten-Be-derived paleo-erosion rates from one-point-eight to three-point-eight million years ago, compared with the modern rate of zero-point-three-four millimeters per year and four-hundred-thousand-year eccentricity pacing. Panel B proposes a nonlinear threshold mechanism: precession-driven South American Monsoon precipitation varies every twenty-three thousand years, but erosion and landsliding respond strongly only once transport conditions exceed a hydrologic threshold, producing large, four-hundred-thousand-year-scale variations.

The preferred explanation for the four-hundred-thousand-year erosion pattern involves nonlinear threshold hillslope behavior driven by landsliding coupled to channel incision. Under this mechanism, erosion rates nonlinearly track precipitation amplitudes once precipitation rises above a threshold.

The proposed mechanism is compelling and well documented in the modern study region, but other explanations remain possible. One alternative is that Pliocene moisture transport was driven directly by four-hundred-thousand-year eccentricity cycles rather than by eccentricity modulating precession amplitudes.

That alternative conflicts with the precession-dominated CLIMBER-2 precipitation results and with late-Quaternary Andean records showing higher-frequency precipitation oscillations driven largely by precession or one-hundred-thousand-year eccentricity cycles. The proposed summary is that Pliocene precipitation was controlled by the strength of the South American Monsoon System, which oscillated on a twenty-three-thousand-year timescale with amplitude controlled by longer four-hundred-thousand-year eccentricity cycles.

The Río Iruya watershed responded nonlinearly through landslide dynamics, allowing relatively small precipitation perturbations of plus or minus twenty percent to produce order-of-magnitude erosion-rate changes with four-hundred-thousand-year pacing. Starting at about two point four million years, the paleo-erosion record shows an overall increase in erosion rate and loss of the four-hundred-thousand-year cyclicity.

One possible explanation is watershed reorganization in the Río Iruya headwaters associated with the provenance shift at about two point four million years. Another possibility is changed precipitation as the Eastern Cordillera continued to grow and concentrated precipitation along its eastern range front.

The study notes that it remains unclear why the reorganized catchment would not continue to record precipitation cyclicity after faithfully tracking four-hundred-thousand-year cyclicity for about one point five million years. The Río Iruya record links erosion cycles to climate across multiple orbital cycles, with modest precipitation changes producing nonlinear erosion responses.

That matters because landscapes may amplify climate signals rather than simply track rainfall directly.